In this study, we comprehensively assessed visual functions, functional vision, and bimanual function in children with uCP to achieve a better understanding of their relation. We found low to moderate correlations between stereoacuity, visual perception, functional vision and bimanual function. Additionally, among visual assessments, visual perception (TVPS-4) was the main predictor of bimanual coordination, bimanual dexterity, and functional hand use with tiny to small effect sizes.
Our results suggest that different aspects of VI play a role in bimanual function. Overall, we found no relation between stereoacuity and bimanual function, except for one weak correlation with bimanual dexterity (TPT). Also in the regression models, stereoacuity only showed null to tiny effect sizes. This is in line with a previous study showing that differences in fine motor skill performance were not predicted by the level of stereoacuity in children with amblyopia (Webber et al., 2008). Hence, our study demonstrates for the first time that also in children with uCP, stereoacuity plays a subordinate role in bimanual function.
Visual perception was mostly correlated with bimanual function. Lower scores on the TVPS-4 subtests were related to lower bimanual coordination, bimanual dexterity, and functional hand use (Kinarm circuit task, TPT, AHA, and CHEQ). This is in line with a previous study in children with uCP (James et al., 2015) reporting that impaired visual perception, assessed with the TVPS-3, was related to reduced quality of motor and processing abilities in daily living, measured with the Assessment of Motor and Process Skills. In our study, we showed for the first time that visual perception was also correlated with bimanual dexterity and bimanual coordination tasks. Longer time to perform a bimanual dexterity task (TPT) and worse bimanual coordination (Kinarm circuit task) were correlated with lower scores on almost all the TVPS-4 subtests. No correlation was found with the other bimanual coordination assessments (Kinarm ball-on-bar level 2 and Box opening task). A possible explanation is that the Kinarm circuit task requires more cognitive demand and the finest and more complex integration of visual stimuli (i.e., recognition and stabilization of the cursor position and keep the ball within the circuit borders) (Decraene et al., 2023), which is not crucial for less complex bimanual coordination tasks such as opening a box and pushing a button (Box opening task) or moving a ball to a fixed target position (Kinarm ball-on-bar level 2). Additionally, our results might suggest that the Box opening task and the Kinarm ball-on-bar level 2 could be more appropriate assessments than the Kinarm circuit and the TPT for evaluating purely bimanual coordination in children with uCP. Furthermore, lower bimanual performance (AHA) was correlated with a lower score on the TVPS-4 subtest visual figure-ground, which was the only subtest that also correlated with all the subscales of perceived quality of bimanual function (CHEQ). Our results support the findings of a previous study, showing that lower scores on the subtests of the Motor-Free Visual Perception Test were related to impaired activities of daily living in children with developmental disabilities (Elbasan et al., 2011).
Overall, our study highlighted that visual figure-ground was the visual perception subtest mostly related to bimanual function. This is in line with one previous study in adults with hemiplegia due to stroke reporting that figure-ground discrimination was the visual perception subtest mostly correlated with an activity of daily living such as putting on and front-fastening a shirt (Mitcham, 1982). In our study, this relation was further confirmed by the elastic-net regression analysis, in which visual figure-ground was the variable showing the most and strongest effect sizes in predicting bimanual function in children with uCP. Our findings could be explained by the organization of the visual system in the brain, involving the ventral and dorsal stream. The dorsal pathway is considered to be responsible for figure-ground processes (Appelbaum et al., 2008) and the processing of visual information for movement control, also known as vision for action, while the ventral pathway is responsible for objects’ recognition, namely vision for perception (Hesse et al., 2012). Hence, visual figure-ground and bimanual function might be controlled by overlapping neural areas, whose damage might impair both visual and bimanual functions in children with uCP. Our results should be considered with caution since estimated effect sizes of the regression models were small. Nevertheless, they might indicate that visual figure-ground could be the visual perception skill to prioritize during assessment of visual function in children with uCP.
Remarkably, visuomotor integration, measured with the subtest of the Beery-VMI, showed a weak association with bimanual function. Lower scores on the VMI subtest were only correlated with longer time taken to perform dexterous bimanual movements (TPT). Additionally, in the regression models, the strongest contribution of the subtest VMI was a very small effect size in predicting the outcomes of the AHA. Since VMI assesses the integration of visual and motor function, we would expect more and stronger relations between this subtest and bimanual function. Nevertheless, it is important to notice that the VMI subtest of the Beery-VMI assesses the ability to copy and draw figures with the dominant hand. Hence, this subtest does not take into account the motor impairments of the non-dominant hand, which largely determines bimanual function in children with uCP (Klingels et al., 2012), potentially explaining the weak associations found in our results.
Interestingly, bimanual dexterity was the only bimanual function significantly correlated with all visual functions (stereoacuity, visual perception, and visuomotor integration). Our results show that bimanual dexterity is the bimanual function for which visual functions are more crucial. Indeed, the TPT assessment entails putting the peg accurately in the hole as fast as possible which requires the highest level of visuomotor integration and eye-hand coordination. Additionally, previous findings suggest that due to impaired stereognosis (Schermann & Tadi, 2024), children with uCP may have to rely more on visual feedback during bimanual dexterity tasks (Decraene et al., 2021). Hence, impairments in visual functions might negatively affect visual feedback, resulting in slower performance on bimanual dexterity tasks.
Functional vision (FCVIQ) was mainly related to perceived quality of bimanual performance (CHEQ-time and feeling). Due to VI in daily life, children with uCP need more time and experience more distress in performing bimanual tasks. This relation was confirmed by the elastic-net regression analysis, in which the FCVIQ total score showed the strongest contribution to the model of the CHEQ-time and CHEQ-feeling. Notably, no relation was found between parent’s reporting on VI in daily life (i.e., FCVIQ) and bimanual dexterity (i.e., TPT), bimanual coordination (i.e., Kinarm and Box opening task), and bimanual performance (i.e., AHA). In our previous study (Crotti et al., 2024), we reported that in our sample of children with uCP, only six children (12%) with data on the FCVIQ have cerebral visual impairment (CVI) while more than 40% of children have some degree of impairment in visual perception. The FCVIQ is a screening questionnaire specifically designed for CVI. Hence, the lack of correlation between the FCVIQ and the bimanual function assessments might be explained by the low variability of our data, since our study included a low number of children with a diagnosis of CVI. On the other hand, the relation between the FCVIQ and the CHEQ could be explained by the fact that both are parent-rated questionnaires. Previous research showed that caregivers often report worse outcomes on questionnaires compared to their children (Robertson et al., 2021; White-Koning et al., 2007). We could hypothesize that parents of children with uCP have the tendency to underestimate the presence of VI of their children due to the diagnosis of the motor impairments which are more prominent and visible in daily life. Nevertheless, no information on the direction (worse or better visual function reported by parents) can be inferred from our analysis and further research is warranted to further understand the specificity of the FCVIQ in detecting VI in children with uCP (Crotti et al., 2024).
Overall regression models did not show a strong performance of visual functions and functional vision in predicting bimanual function. The Kinarm circuit task was the only model showing a large predictive performance. Nevertheless, this result was mainly driven by age, which was the strongest predictor. Our results are not totally unexpected since other factors (e.g., motor and sensorimotor impairments), which were not assessed by the predictors of our models (i.e., visual assessments) have a large impact on bimanual function in children with uCP. Although showing a weak predicting performance, regression models were significant, with visual assessments reporting tiny to small effect sizes in predicting bimanual dexterity, bimanual coordination, and functional hand use. Hence, our results highlight the presence of a relation between visual and bimanual function in children with uCP. Furthermore, additional visual functions (e.g., visual feedback, visual spatial attention), which were not included in our models, could have a potential role in impacting bimanual function and therefore, they should be addressed in future studies in children with uCP (Hawe et al., 2020).
Some limitations of our study should be noted. First, the relatively small sample size could lead to imprecise parameter estimates of the regression models. To overcome the risk of overfitting, we performed an elastic-net regularized regression which allows to handle more predictors compared to the sample size (Zou & Hastie, 2005). Additionally, technical issues with the Box opening task resulted in more missing data for this assessment, which might influence the absence of significant correlations with the visual functions and functional vision assessments. Lastly, the low variance explained by VI supports the need for clinicians to consider additional factors (e.g., stereognosis, cognitive function, visuospatial attention) that may impact bimanual function in children with uCP (Decraene et al., 2021; Swinnen & Gooijers, 2015).